farrokhi/pg_repack
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Fix up commit 4b3347, where I failed to notice that the initial

Browse Source 
problem was really about the OID being interpreted as an integer
literal upon input, and overflowing its integer space before even making
it into pg_try_advisory_lock(). (We do still need to add -2147483648 to
make the result fit into an integer, as 4b3347 does.)

Hopefully fixes issue #30, for real this time.
This commit is contained in:
Josh Kupershmidt 2014-11-20 17:18:15 -05:00
parent 4b334745a3
commit 6d3c085b22
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
bin/pg_repack.c
View File

@ -1359,7 +1359,8 @@ repack_one_table(const repack_table *table, const char *orderby)
/* Release advisory lock on table. */
params[0] = REPACK_LOCK_PREFIX_STR;
params[1] = buffer;
res = pgut_execute(connection, "SELECT pg_advisory_unlock($1, -2147483648 + $2)",
res = pgut_execute(connection, "SELECT pg_advisory_unlock($1, CAST(-2147483648 + $2::bigint AS integer))",
2, params);
ret = true;
@ -1530,7 +1531,7 @@ static bool advisory_lock(PGconn *conn, const char *relid)
* *unsigned* 4-byte integer. Add -2147483648 to that OID to make
* it fit reliably into signed int space.
*/
res = pgut_execute(conn, "SELECT pg_try_advisory_lock($1, -2147483648 + $2)",
res = pgut_execute(conn, "SELECT pg_try_advisory_lock($1, CAST(-2147483648 + $2::bigint AS integer))",
2, params);
if (PQresultStatus(res) != PGRES_TUPLES_OK) {